1. Introduction
Ž .
Since their domestication at least 12,000 years ago Davis and Valla, 1979 , domestic dogs, Canis familiaris have lived in close association with people fulfilling a variety of
roles, many of which require the dog to be trained. Today, the majority of dogs in Western society are kept as companions. Within this capacity the importance of training
and the formation of a successful dog–human relationship remains paramount. Conse- quently, there has amassed a large amount of popular literature advocating methods for
Ž training and modifying relationships between dogs and people e.g., Rogerson, 1992;
. Abrantes, 1997; Appleby, 1997 .
Many of the ideas put forward in these texts are based on ethological concepts relating to intraspecific social behaviour of the dog, and many refer to the behaviour of
the ancestor, the wolf, Canis lupus. There is one extremely common assumption, that dogs behave towards people as though they were conspecifics which have been
Ž .
incorporated within their hierarchy e.g., Blackshaw, 1985 . It is consequently assumed that dog–human interactions have the same social implications as their dog–dog
Ž .
counterparts. This opinion is summed up by Hediger 1965, p. 295 who states that ‘‘there is no doubt about the fact that the dog considers its human master as a socially
superior member of its own species — as a member of its pack’’. This lack of doubt is surprising given the absence of empirical data on this subject. Extrapolation from
dog–dog behaviour to the dog–human context occurs with reference to all aspects of social behaviour, and is particularly common when discussing play.
The domestic dog is renowned for exhibiting high levels of play. In addition to Ž
. intraspecific play, which is ubiquitous amongst mammals Suomi, 1982 , the dog is one
Ž .
of the few species to engage in interspecific play Russell, 1936 . Play occupies a large Ž
. proportion of the interaction time between dog and owner Hart, 1995 , yet has received
Ž .
very little scientific study Nott, 1992 . Common opinion is that interspecific play is equivalent to intraspecific play and general acceptance of this homogeneity has led to
ideas about the effects of dog–dog play being applied to dog–human play. Assumptions of homogeneity take two forms. Firstly, it is claimed that dog–human
play can serve as compensation for a deficiency in dog–dog play, e.g., ‘‘when no other Ž
dog is available then the owner can fill the need of being a playmate’’ Whitney, 1989, .
p. 98 , ‘‘if deprived of the opportunity to play with other dogs, they have no other Ž
. choice but to direct play behaviour towards humans’’ Askew, 1996, p. 171 and ‘‘a dog
that lives with another dog will usually play more games with that dog than with its Ž
. owner’’ Rogerson, 1992, p. 55 . These authors suggest that opportunities for intraspe-
cific play reduce the performance of interspecific play, thereby implying the two are motivationally homogeneous, but we know of no data to support this claim. Secondly,
Ž .
there are observations about the structure of play. Aldis 1975 states that dog–human play is composed of the same behavioural patterns as dog–dog play. Mitchell and
Ž .
Thompson 1991a compared their own observations of dog–human play to others’ Ž
. e.g., Aldis, 1975 of dog–dog play. They noted the performance of only two additional
1
Ž ‘‘projects’’
by dogs when playing with humans one of which was in response to a
1
‘‘Projects’’ are defined as ‘‘sequences of actions that are repeated in order to calibrate an organism’s Ž
. control over these actions or the actions of the play partner’’ Mitchell and Thompson, 1991a, p. 189 .
. person pretending to throw a ball, which a dog would not be capable of doing . On this
basis they describe the two play types as similar. However, the observation that similar actions occur when playing with other dogs and with humans is not evidence that the
structure is the same, since identical actions can be combined in different ways to fulfil different functions. There exists no quantitative data comparing the structure of dog–dog
and dog–human play.
In this paper we examine the hypothesis that dog–human and dog–dog play are not homogeneous categories. To do so we examine both quantity and quality of play. We
present two investigations which test the claim that dogs living with other dogs play less frequently with their owner than dogs that live alone. The first investigation involved
focal sampling in dog walking areas and compared the incidence of play in one and multi-dog partnerships. In the second we incorporated questions within a national survey
of dog owners, and examined reported dog–human play frequencies with single and multiple-housed dogs. The initial hypothesis would predict a higher incidence of
interspecific play in single-housed dogs.
We then test the null hypothesis that the structure of dog–dog and dog–human play are the same, even when the actions compared are restricted to those occurring in
intraspecific play. In an experimental study of 12 Labrador Retrievers, we compared both the composition and the competitiveness of intra- and interspecific object-oriented
play.
2. Focal sampling
Kategori